recreate media types file

scripts/generate_default_media_types.py

@@ -0,0 +1,98 @@
+# pragma: no cover
+from __future__ import annotations
+
+import argparse
+import http.client
+import re
+from contextlib import closing
+from pathlib import Path
+
+module_dir = Path(__file__).parent.resolve()
+media_types_py = module_dir / "../src/servestatic/media_types.py"
+
+
+def main() -> int:
+    parser = argparse.ArgumentParser()
+    parser.add_argument("--check", action="store_true")
+    args = parser.parse_args()
+
+    func_str = get_default_types_function()
+    text = media_types_py.read_text()
+    new_text = re.sub(
+        r"def default_types.*\}",
+        func_str,
+        text,
+        flags=re.DOTALL,
+    )
+    if new_text != text:
+        if args.check:
+            print("Would write changes")
+            return 1
+        print(f"Writing {media_types_py}")
+        media_types_py.write_text(new_text)
+    return 0
+
+
+EXTRA_MIMETYPES = {
+    # Nginx uses application/javascript, but HTML specification recommends text/javascript:
+    ".js": "text/javascript",
+    ".md": "text/markdown",
+    ".mjs": "text/javascript",
+    ".woff": "application/font-woff",
+    ".woff2": "font/woff2",
+    "apple-app-site-association": "application/pkc7-mime",
+    # Adobe: https://www.adobe.com/devnet-docs/acrobatetk/tools/AppSec/xdomain.html#policy-file-host-basics
+    "crossdomain.xml": "text/x-cross-domain-policy",
+}
+
+
+FUNCTION_TEMPLATE = '''\
+def default_types() -> dict[str, str]:
+    """
+    We use our own set of default media types rather than the system-supplied
+    ones. This ensures consistent media type behaviour across varied
+    environments.  The defaults are based on those shipped with nginx, with
+    some custom additions.
+
+    (Auto-generated by scripts/generate_default_media_types.py)
+    """
+    return {{
+{entries}
+    }}'''
+
+
+def get_default_types_function() -> str:
+    types_map = get_types_map()
+    lines = [f'        "{suffix}": "{media_type}",' for suffix, media_type in types_map.items()]
+    return FUNCTION_TEMPLATE.format(entries="\n".join(lines))
+
+
+def get_types_map() -> dict[str, str]:
+    nginx_data = get_nginx_data()
+    matches = re.findall(r"(\w+/.*?)\s+(.*?);", nginx_data)
+    types_map = {}
+    for match in matches:
+        media_type = match[0]
+        # This is the default media type anyway, no point specifying it explicitly
+        if media_type == "application/octet-stream":
+            continue
+
+        extensions = match[1].split()
+        for extension in extensions:
+            types_map[f".{extension}"] = media_type
+    types_map.update(EXTRA_MIMETYPES)
+    return dict(sorted(types_map.items()))
+
+
+def get_nginx_data() -> str:
+    conn = http.client.HTTPSConnection("raw.githubusercontent.com")
+    with closing(conn):
+        conn.request("GET", "/nginx/nginx/master/conf/mime.types")
+        response = conn.getresponse()
+        if response.status != 200:
+            raise AssertionError
+        return response.read().decode()
+
+
+if __name__ == "__main__":
+    raise SystemExit(main())